Visualizing and Maintaining the Green Canopy of NYC

STA 9750 Mini-Project #03

Author

Sabrina Zhu

Published

November 16, 2025

NoteProject Overview


Data Period: 2015 - 2024

Data Source:

City Council Districts | NYC Forestry Tree Points | Tree Safety Risk | Tree Maintenance Orders

Project Structure:

  • Data Acquisition — Downloading tree, district, risk, and maintenance data
  • Data Integration — Spatial joining and processing
  • Initial Explorations — District-level tree analysis
  • Government Project — District 2 tree health initiative proposal


Context

Trees play a crucial role in making New York City livable and sustainable. Nearly 900,000 street trees, maintained by the NYC Parks Department and community partners, provide environmental and social benefits to millions of residents. Using comprehensive data on tree health, species, and coverage across neighborhoods, this analysis identifies patterns and gaps to inform targeted urban forestry programs that ensure every district has access to healthy, thriving green spaces.


Data Aquisition

To explore tree distribution and health across NYC, spatial and environmental data were gathered from official NYC sources using “polite” downloading practices that respect data providers through caching and rate limiting.

Show the code 
# Load libraries
library(sf)
library(tidyverse)
library(leaflet)
library(DT)
library(data.table)

NYC City Council District Boundaries

NYC is divided into 51 City Council districts, and analyzing tree patterns by district requires accurate boundary data. The official district shapefile was downloaded from the NYC Department of Planning, providing the geographic outlines needed to link trees to the correct district.

Show the code 
download_nycc_districts <- function() {
  # Create directory if needed
  if (!dir.exists(file.path("data", "mp03"))) {
    dir.create(file.path("data", "mp03"), showWarnings = FALSE, recursive = TRUE)
  }
  
  # File paths
  zip_file <- file.path("data", "mp03", "nycc_25c.zip")
  shp_file <- file.path("data", "mp03", "nycc_25c", "nycc.shp")
  
  # Download and unzip if needed
  if (!file.exists(shp_file)) {
    if (!file.exists(zip_file)) {
      download.file(
        "https://s-media.nyc.gov/agencies/dcp/assets/files/zip/data-tools/bytes/city-council/nycc_25c.zip",
        zip_file, mode = "wb", quiet = TRUE
      )
    }
    unzip(zip_file, exdir = file.path("data", "mp03"))
  }
  
  # Read and transform
  districts <- st_read(shp_file, quiet = TRUE) |>
    st_transform("WGS84") |>
    st_make_valid()
  
  return(districts)
}

Tree Points

The NYC Forestry Tree Points dataset from NYC OpenData was used, which lists every recorded street tree in the city along with details like location, species, and condition. This dataset provides a complete view of the urban forest and supports comparison of tree patterns across districts.

Show the code 
download_trees <- function(sample_size = NULL) {
  cache_file <- "data/mp03/all_trees.rds"

  # Check cache first
  if (file.exists(cache_file)) {
    trees <- readRDS(cache_file)
    
    # Sample if requested
    if (!is.null(sample_size)) {
      set.seed(123)
      trees <- trees |> slice_sample(n = min(sample_size, nrow(trees)))
    }
    return(trees)
  }
  
  # Download from API using GeoJSON
  base_url <- "https://data.cityofnewyork.us/resource/hn5i-inap.geojson"
  all_trees <- list()
  offset <- 0
  limit <- 50000
  
  repeat {
    url <- paste0(base_url, "?$limit=", limit, "&$offset=", format(offset, scientific = FALSE))
    temp_file <- tempfile(fileext = ".geojson")
    
    tryCatch({
      download.file(url, temp_file, mode = "wb", quiet = TRUE)
      chunk <- st_read(temp_file, quiet = TRUE)
      
      if (nrow(chunk) == 0) break
      
      all_trees[[length(all_trees) + 1]] <- chunk
      
      if (nrow(chunk) < limit) break
      offset <- offset + limit
    }, error = function(e) {
      break
    })
  }
  
  # Combine and process
  cat("Combining and processing tree data...\n")
  trees <- bind_rows(all_trees) |>
    st_make_valid() |>
    select(tree_id, spc_common, spc_latin, tree_dbh, health, status, cncldist, geometry) |>
    mutate(
      tpcondition = factor(
        ifelse(is.na(health), "Unknown", health),
        levels = c("Excellent", "Good", "Fair", "Poor", "Critical", "Dead", "Unknown")
      ),
      genusspecies = as.factor(spc_latin)
    )
  
  # Save to cache
  saveRDS(trees, cache_file)
  
  # Sample if requested
  if (!is.null(sample_size)) {
    set.seed(123)
    trees <- trees |> slice_sample(n = min(sample_size, nrow(trees)))
  }
  
  return(trees)
}

Tree Safety Risk Data

The Parks Department’s Tree Safety Risk Assessment dataset documents trees that pose potential safety hazards.

Show the code 
download_tree_risk <- function() {
  if (!dir.exists(file.path("data", "mp03"))) {
    dir.create(file.path("data", "mp03"), showWarnings = FALSE, recursive = TRUE)
  }
  
  cache_file <- file.path("data", "mp03", "tree_risk.rds")
  
  # Check cache
  if (file.exists(cache_file)) {
    cat("Loading cached tree risk assessment data...\n")
    risk_data <- readRDS(cache_file)
    cat("✓ Loaded", format(nrow(risk_data), big.mark = ","), "risk records from cache\n\n")
    return(risk_data)
  }
  
  # Download from API using JSON
  cat("Downloading tree safety risk assessment data...\n")
  base_url <- "https://data.cityofnewyork.us/resource/259a-b6s7.json"
  all_data <- list()
  offset <- 0
  limit <- 50000
  
  repeat {
    url <- paste0(base_url, "?$limit=", limit, "&$offset=", format(offset, scientific = FALSE))
    
    tryCatch({
      chunk <- jsonlite::fromJSON(url)
      
      if (nrow(chunk) == 0) break
      
      all_data[[length(all_data) + 1]] <- chunk
      cat("  Downloaded", format(offset + nrow(chunk), big.mark = ","), "risk records...\n")
      
      if (nrow(chunk) < limit) break
      offset <- offset + limit
    }, error = function(e) {
      cat("  Error at offset", offset, ":", e$message, "\n")
      break
    })
  }
  
  # Combine all data
  risk_data <- bind_rows(all_data)
  
  # Save to cache
  saveRDS(risk_data, cache_file)
  cat("✓ Risk data cached for future use\n")
  cat("✓ Total risk records downloaded:", format(nrow(risk_data), big.mark = ","), "\n\n")
  
  return(risk_data)
}

Tree Maintenance Orders

The Tree Maintenance Orders dataset tracks ongoing and pending maintenance work orders for NYC street trees.

Show the code 
download_tree_maintenance <- function() {
  if (!dir.exists(file.path("data", "mp03"))) {
    dir.create(file.path("data", "mp03"), showWarnings = FALSE, recursive = TRUE)
  }
  
  cache_file <- file.path("data", "mp03", "tree_maintenance.rds")
  
  # Check cache
  if (file.exists(cache_file)) {
    cat("Loading cached tree maintenance data...\n")
    maintenance_data <- readRDS(cache_file)
    cat("✓ Loaded", format(nrow(maintenance_data), big.mark = ","), "maintenance records from cache\n\n")
    return(maintenance_data)
  }
  
  # Download from API using JSON
  cat("Downloading tree maintenance orders data...\n")
  base_url <- "https://data.cityofnewyork.us/resource/bdjm-n7q4.json"
  all_data <- list()
  offset <- 0
  limit <- 50000
  
  repeat {
    url <- paste0(base_url, "?$limit=", limit, "&$offset=", format(offset, scientific = FALSE))
    
    tryCatch({
      chunk <- jsonlite::fromJSON(url)
      
      if (nrow(chunk) == 0) break
      
      all_data[[length(all_data) + 1]] <- chunk
      cat("  Downloaded", format(offset + nrow(chunk), big.mark = ","), "maintenance records...\n")
      
      if (nrow(chunk) < limit) break
      offset <- offset + limit
    }, error = function(e) {
      cat("  Error at offset", offset, ":", e$message, "\n")
      break
    })
  }
  
  # Combine all data
  maintenance_data <- bind_rows(all_data)
  
  # Save to cache
  saveRDS(maintenance_data, cache_file)
  cat("✓ Maintenance data cached for future use\n")
  cat("✓ Total maintenance records downloaded:", format(nrow(maintenance_data), big.mark = ","), "\n\n")
  
  return(maintenance_data)
}


Data Integration and Initial Exploration

Show the code 
# Optimized spatial join with chunking
spatial_join <- function(trees, districts) {
  cache_file <- "data/mp03/trees_joined.rds"
  
  if (file.exists(cache_file)) {
    result <- readRDS(cache_file)
    return(result)
  }
  
  # Process in 100k chunks
  n <- nrow(trees)
  chunk_size <- 100000
  chunks <- split(seq_len(n), ceiling(seq_len(n) / chunk_size))
  
  result <- map_dfr(chunks, ~{
    chunk_result <- st_join(trees[.x, ], districts, join = st_intersects, left = TRUE)
    chunk_result
  })
  
  saveRDS(result, cache_file)
  result
}

# Load all data
districts <- download_nycc_districts()
trees <- download_trees(sample_size = NULL)
trees_joined <- spatial_join(trees, districts)

# Convert to data.table for fast analysis
trees_dt <- trees_joined |> st_drop_geometry() |> as.data.table()

Mapping NYC Trees

After pulling the datasets, I mapped a sample of 50,000 trees to get a quick look at how trees are spread across the city without slowing down the visualization. Even though it’s only a slice of the full data, it still shows the overall pattern across NYC. You can follow the color scale or hover over each district to see the actual tree counts for a closer comparison.

Show the code 
library(leaflet.extras)

# Sample trees
set.seed(123)
trees_map_sample <- trees_joined |> slice_sample(n = 50000)

# Calculate district summaries
district_summary <- trees_dt[!is.na(CounDist), .N, by = CounDist]
districts_map <- districts |> left_join(district_summary, by = "CounDist")

# Color palette for districts
pal_districts <- colorNumeric("Greens", district_summary$N)

# Define colors
condition_colors <- c(
  "Excellent" = "#006400",
  "Good" = "#228B22", 
  "Fair" = "#FFD700",
  "Poor" = "#FF8C00",
  "Critical" = "#DC143C",
  "Dead" = "#8B0000",
  "Unknown" = "#808080"
)

# Extract coordinates and add color column BEFORE dropping geometry
coords <- st_coordinates(trees_map_sample)
trees_map_sample$lon <- coords[,1]
trees_map_sample$lat <- coords[,2]
trees_map_sample$tree_color <- condition_colors[as.character(trees_map_sample$tpcondition)]

# Now drop geometry for leaflet
trees_df <- trees_map_sample |> st_drop_geometry()

# Create map with layer groups for filtering
leaflet() |>
  addProviderTiles(providers$CartoDB.Positron) |>
  
  # Districts as a toggleable layer
  addPolygons(
    data = districts_map,
    fillColor = ~pal_districts(N),
    fillOpacity = 0.4,
    color = "#CCCCCC",  
    weight = 2,
    label = ~paste0("District ", CounDist, ": ", format(N, big.mark = ","), " trees"),
    highlightOptions = highlightOptions(
      weight = 3,
      color = "#999999",    
      fillOpacity = 0.8,  
      bringToFront = FALSE
    ),
    group = "Districts"  # Add to Districts group
  ) |>
  
  # Trees as a toggleable layer
  addCircles(
    data = trees_df,
    lng = ~lon,
    lat = ~lat,
    radius = 3,
    color = ~tree_color,
    fillColor = ~tree_color,
    weight = 1,
    opacity = 0.7,
    fillOpacity = 0.4,
    group = "Trees"  # Add to Trees group
  ) |>
  
  # Add Layer Control to toggle trees/districts
  addLayersControl(
    overlayGroups = c("Districts", "Trees"),
    options = layersControlOptions(collapsed = FALSE),
    position = "topright"
  ) |>
  
  # Tree Condition Legend - TOP LEFT
  addLegend(
    position = "topleft",
    colors = condition_colors,
    labels = names(condition_colors),
    title = "Tree Condition",
    opacity = 0.7
  ) |>
  
  # Trees per District Legend - BOTTOM RIGHT
  addLegend(
    position = "bottomright", 
    pal = pal_districts, 
    values = district_summary$N, 
    title = "Trees per District"
  ) |>
  
  # Tree count info - BOTTOM LEFT
  addControl(
    html = paste0(
      "<b>Showing 50,000 of ", format(nrow(trees_joined), big.mark = ","), " trees</b>",
      "</div>"
    ),
    position = "bottomleft"
  ) |>
  
  setView(lng = -73.95, lat = 40.7, zoom = 11)


District-Level Analysis of Tree Coverage

Now it’s time to explore the data by joining the tree points with the district boundaries and answering the following questions.

1. Total Tree Count by Council District

Show the code 
q1 <- trees_dt[!is.na(CounDist), .N, by = CounDist][order(-N)] |>
  mutate(Rank = row_number()) |>
  select(Rank, CounDist, N) |>
  setnames(c("CounDist", "N"), c("Council District", "Number of Trees"))

datatable(q1, options = list(pageLength = 10, dom = 'tp'), rownames = FALSE) |>
  formatStyle('Rank', target = 'row', backgroundColor = styleEqual(1, '#dbe5ff')) |>
  formatCurrency('Number of Trees', currency = "", interval = 3, mark = ",", digits = 0)


2. Tree Density Across Council Districts

District 7 achieves the highest density at 0.2835 trees per 1000 square units.

Show the code 
q2 <- trees_dt[!is.na(CounDist), .N, by = CounDist] |>
  left_join(districts |> st_drop_geometry() |> select(CounDist, Shape_Area), by = "CounDist") |>
  mutate(tree_density = (N / Shape_Area) * 1000) |>
  arrange(desc(tree_density)) |>
  mutate(Rank = row_number()) |>
  select(Rank, CounDist, N, Shape_Area, tree_density) |>
  setnames(c("CounDist", "N", "Shape_Area", "tree_density"), 
           c("Council District", "Number of Trees", "Area (sq units)", "Tree Density"))

datatable(q2, options = list(pageLength = 10, dom = 'tp'), rownames = FALSE) |>
  formatStyle('Rank', target = 'row', backgroundColor = styleEqual(1, '#ffe4b3')) |>
  formatCurrency('Number of Trees', currency = "", interval = 3, mark = ",", digits = 0) |>
  formatRound(c('Area (sq units)', 'Tree Density'), c(2, 4))


3. Districts with the Highest Share of Dead Trees

Show the code 
q3 <- trees_dt[!is.na(CounDist), .(
  total_trees = .N,
  dead_trees = sum(tpcondition == "Dead", na.rm = TRUE)
), by = CounDist][, percent_dead := (dead_trees / total_trees) * 100][order(-percent_dead)] |>
  mutate(Rank = row_number()) |>
  select(Rank, CounDist, total_trees, dead_trees, percent_dead) |>
  setnames(c("CounDist", "total_trees", "dead_trees", "percent_dead"),
           c("Council District", "Total Trees", "Dead Trees", "Percent Dead"))

datatable(q3, options = list(pageLength = 10, dom = 'tip'), rownames = FALSE) |>
  formatStyle('Rank', target = 'row', backgroundColor = styleEqual(1, '#ffcccc')) |>
  formatCurrency(c('Total Trees', 'Dead Trees'), currency = "", interval = 3, mark = ",", digits = 0) |>
  formatRound('Percent Dead', digits = 2)


4. Most Common Tree Species in Manhattan

Show the code 
trees_dt[, Borough := fcase(
  between(CounDist, 1, 10), "Manhattan",
  between(CounDist, 11, 18), "Bronx",
  between(CounDist, 19, 32), "Queens",
  between(CounDist, 33, 48), "Brooklyn",
  between(CounDist, 49, 51), "Staten Island"
)]

q4 <- trees_dt[Borough == "Manhattan" & !is.na(genusspecies), .N, by = genusspecies][order(-N)] |>
  mutate(Rank = row_number(), percent = N / sum(N) * 100) |>
  select(Rank, genusspecies, N, percent) |>
  setnames(c("genusspecies", "N", "percent"), c("Species", "Count", "Percent of Total"))

datatable(q4, options = list(pageLength = 10, dom = 'tip'), rownames = FALSE) |>
  formatStyle('Rank', target = 'row', backgroundColor = styleEqual(1, '#c8e6c9')) |>
  formatCurrency('Count', currency = "", interval = 3, mark = ",", digits = 0) |>
  formatRound('Percent of Total', digits = 2)


5. Tree Species Nearest to Baruch College

Show the code 
baruch <- st_sfc(st_point(c(-73.9834, 40.7403)), crs = "WGS84")

q5 <- trees_joined |>
  mutate(distance_ft = as.numeric(st_distance(geometry, baruch)) * 3.28084) |>
  arrange(distance_ft) |>
  slice(1:10) |>
  st_drop_geometry() |>
  mutate(Rank = row_number()) |>
  select(Rank, genusspecies, distance_ft, tpcondition, CounDist) |>
  setnames(c("genusspecies", "distance_ft", "tpcondition", "CounDist"),
           c("Species", "Distance (ft)", "Condition", "Council District"))

datatable(q5, options = list(pageLength = 10, dom = 't'), rownames = FALSE) |>
  formatStyle('Rank', target = 'row', backgroundColor = styleEqual(1, '#fff4cc')) |>
  formatRound('Distance (ft)', digits = 0)


Government Project Design

Proposal: District 2 Street Tree Revitalization Program

District 2 is home to some of the most walked streets in the city, including the area around Baruch College. Trees here play a big role in keeping these busy neighborhoods comfortable, shaded, and lively. But the data shows that many trees in the district are aging or in poor condition, and the canopy is thinner than it should be for such a dense and active area. This proposal suggests a Tree Revitalization Program focused on replacing dead trees, improving tree health, and strengthening the canopy for the long term.

Show the code 
# Comprehensive district analysis with safety and maintenance data
dead_analysis <- trees_dt[!is.na(CounDist), .(
  total_trees = .N,
  dead_trees = sum(tpcondition == "Dead", na.rm = TRUE),
  critical_trees = sum(tpcondition == "Critical", na.rm = TRUE),
  poor_trees = sum(tpcondition == "Poor", na.rm = TRUE),
  unhealthy_trees = sum(tpcondition %in% c("Dead", "Critical", "Poor"), na.rm = TRUE),
  excellent_trees = sum(tpcondition == "Excellent", na.rm = TRUE),
  good_trees = sum(tpcondition == "Good", na.rm = TRUE)
), by = CounDist][, `:=`(
  percent_dead = (dead_trees / total_trees) * 100,
  percent_unhealthy = (unhealthy_trees / total_trees) * 100
)][order(-percent_dead)]

# Add borough classification
dead_analysis[, Borough := fcase(
  between(CounDist, 1, 10), "Manhattan",
  between(CounDist, 11, 18), "Bronx",
  between(CounDist, 19, 32), "Queens",
  between(CounDist, 33, 48), "Brooklyn",
  between(CounDist, 49, 51), "Staten Island"
)]

# District 2 focus
d2 <- dead_analysis[CounDist == 2]
manhattan_districts <- dead_analysis[Borough == "Manhattan"][order(-percent_dead)]
healthiest_manhattan <- manhattan_districts[CounDist != 2][order(percent_dead)][1]
top5_city <- dead_analysis[1:5]
NoteDistrict 2 Key Metrics


🌳 Total Trees: 11,562

💀 Dead Trees: 1,576 (13.63% of total)

⚠️ Critical Condition: 61 (0.53% of total)

⚡ Poor Condition: 291 (2.52% of total)

❌ Total Unhealthy: 1,928 (16.68% of total)


Zoomed-In Map: District 2 Tree Conditions

Show the code 
d2_boundary <- districts |> filter(CounDist == 2)
d2_trees <- trees_joined |> filter(CounDist == 2)

# Color palette
pal_d2 <- colorFactor(
  palette = c("#006400", "#228B22", "#FFD700", "#FF8C00", "#DC143C", "#8B0000", "#808080"),
  levels = c("Excellent", "Good", "Fair", "Poor", "Critical", "Dead", "Unknown")
)

leaflet() |>
  addProviderTiles(providers$CartoDB.Positron) |>
  addPolygons(
    data = d2_boundary,
    fillColor = "transparent",
    color = "#000000",
    weight = 3,
    label = "District 2"
  ) |>
  addCircleMarkers(
    data = d2_trees,
    radius = 3,
    color = ~pal_d2(tpcondition),
    fillColor = ~pal_d2(tpcondition),
    fillOpacity = 0.8,
    stroke = TRUE,
    weight = 1,
    label = ~paste0(
      genusspecies, " - ", 
      tpcondition
    ),
    labelOptions = labelOptions(
      style = list("font-weight" = "normal", "padding" = "3px 8px"),
      textsize = "12px",
      direction = "auto"
    ),
    popup = ~paste0(
      "<b>", genusspecies, "</b><br>",
      "Condition: ", tpcondition, "<br>",
      "Diameter: ", dbh, " inches"
    )
  ) |>
  addLegend(
    "topright",
    pal = pal_d2,
    values = d2_trees$tpcondition,
    title = "Tree Condition",
    opacity = 1
  ) |>
  addControl(
    paste0("<b>District 2: ", format(nrow(d2_trees), big.mark = ","), " trees</b>"),
    "bottomleft"
  )


The map above reveals concentrated clusters of dead and critical trees, particularly in the eastern portions of District 2, indicating areas where urgent intervention is needed.


Why District 2?

Quantitative Comparison

District 2 stands out because it has one of the highest percentages of dead trees in the area, as found in question 3 of our initial exploration.

Show the code 
comparison <- top5_city |>
  mutate(Rank = row_number()) |>
  select(Rank, CounDist, Borough, total_trees, dead_trees, unhealthy_trees, percent_dead, percent_unhealthy) |>
  setnames(c("CounDist", "Borough", "total_trees", "dead_trees", "unhealthy_trees", "percent_dead", "percent_unhealthy"),
           c("District", "Borough", "Total Trees", "Dead", "Unhealthy", "% Dead", "% Unhealthy"))

datatable(
  comparison,
  options = list(pageLength = 5, dom = 't'),
  rownames = FALSE,
  caption = "Top 5 NYC Districts by Tree Mortality Rate"
) |>
  formatStyle(
    'Rank', 
    target = 'row', 
    backgroundColor = styleEqual(which(comparison$District == 2), '#fff4cc')
  ) |>
  formatCurrency(c('Total Trees', 'Dead', 'Unhealthy'), currency = "", interval = 3, mark = ",", digits = 0) |>
  formatRound(c('% Dead', '% Unhealthy'), digits = 2)


As shown above, District 2 ranks #3 citywide for tree mortality, but within Manhattan it tells a more concerning story. District 2 ranks #1 in Manhattan for tree mortality with a 13.63% dead tree rate, the highest among all 10 Manhattan districts. While District 2 has a relatively modest tree inventory of 11,562 trees (smaller than districts 8, 10, and 9), it faces disproportionately severe health challenges. The district’s 1,637 safety risk trees (dead + critical) represent 14.16% of its total inventory, the highest safety risk percentage in Manhattan and 1.76 percentage points higher than the second-worst district (District 1 at 12.40%).

Show the code 
# Manhattan comparison with safety metrics - Light red highlight
manhattan_comp_enhanced <- manhattan_districts |>
  mutate(
    Rank = row_number(),
    safety_risk = dead_trees + critical_trees,
    pct_safety_risk = (safety_risk / total_trees) * 100
  ) |>
  select(
    Rank, CounDist, total_trees, dead_trees, critical_trees, 
    unhealthy_trees, safety_risk, percent_dead, percent_unhealthy, pct_safety_risk
  ) |>
  setnames(
    c("CounDist", "total_trees", "dead_trees", "critical_trees", 
      "unhealthy_trees", "safety_risk", "percent_dead", "percent_unhealthy", "pct_safety_risk"),
    c("District", "Total Trees", "Dead", "Critical", 
      "Unhealthy", "Safety Risk", "% Dead", "% Unhealthy", "% Safety Risk")
  )

datatable(
  manhattan_comp_enhanced,
  options = list(pageLength = 10, dom = 't', scrollX = TRUE),
  rownames = FALSE,
  caption = "Manhattan Districts - Tree Health & Safety Analysis"
) |>
  formatStyle(
    'Rank',
    target = 'row',
    backgroundColor = styleEqual(which(manhattan_comp_enhanced$District == 2), '#ffe6e6')  # Light red
  ) |>
  formatCurrency(c('Total Trees', 'Dead', 'Critical', 'Unhealthy', 'Safety Risk'), 
                 currency = "", interval = 3, mark = ",", digits = 0) |>
  formatRound(c('% Dead', '% Unhealthy', '% Safety Risk'), digits = 2)


Current Maintenance Backlog

Show the code 
download_tree_maintenance_d2 <- function() {
  cache_file <- "data/mp03/tree_maintenance_d2.rds"
  
  if (file.exists(cache_file)) {
    return(readRDS(cache_file))
  }
  
  # Download District 2 maintenance data
  url <- "https://data.cityofnewyork.us/resource/bdjm-n7q4.json?$limit=50000"
  maintenance <- jsonlite::fromJSON(url)
  
  saveRDS(maintenance, cache_file)
  return(maintenance)
}

# Add this with your other data loading
d2_maintenance <- download_tree_maintenance_d2()

District 2’s maintenance system reveals significant gaps between need and action:

WarningMaintenance Gap Analysis


Trees Requiring Urgent Attention:

  • 💀 1,576 dead trees requiring immediate removal

  • ⚠️ 61 critical condition trees needing urgent treatment

  • Total urgent needs: 1,637 trees

Actual Maintenance Orders on Record:

  • 📋 Only 1,109 active work orders for District 2

  • 🪓 66 tree removal orders (covers only 4% of dead trees)

  • 🌳 38 stump removal orders (blocking new plantings)

The Gap: District 2 has 1,637 trees requiring urgent intervention but only 66 removal orders on file. This means 96% of dead trees have not yet been scheduled for removal.


Thanks for Reading!